Dehydrogenation of Simple Hydrocarbons on Platinum Cluster Ions

J. Phys. Chem. A 2002, 106, 11465-11469

11465

Dehydrogenation of Simple Hydrocarbons on Platinum Cluster Ions Tetsu Hanmura,† Masahiko Ichihashi,‡ and Tamotsu Kondow*,‡ East Tokyo Laboratory, Genesis Research Institute, Inc., 717-86 Futamata, Ichikawa, Chiba 272-0001, Japan, and Cluster Research Laboratory, Toyota Technological Institute in East Tokyo Laboratory, Genesis Research Institute, Inc., 717-86 Futamata, Ichikawa, Chiba 272-0001, Japan ReceiVed: May 22, 2002; In Final Form: September 13, 2002

Reactions of size-selected platinum cluster ions, Ptn+ (n ) 1-5), with small hydrocarbon molecules (CH4, C2H6, C2H4, and C2H2) were studied in a beam-gas cell geometry at collision energies less than 1 eV employing a tandem mass spectrometer. It was found that dehydrogenation of the hydrocarbon molecules proceeds exclusively on Ptn+ and the absolute cross sections of the reactions depend critically on the cluster size, n, and the collision energy. The size- and the collision energy-dependent cross sections are successfully explained in an RRK framework.

1. Introduction Platinum is one of the most useful elements in many catalytic reactions such as hydrogenation, dehydrogenation, and cracking of various hydrocarbons.1,2 In practical catalysts, platinum is frequently dispersed on a solid support as fine particles whose reactivity depends critically on their diameters.3 It has been reported, for instance, that the rate of hydrogenation of ethylene on platinum particles increases as the particle diameters decrease down to ∼0.6 nm,4,5 which corresponds to the cluster size (the number of the constituent platinum atoms) of ∼10. There has been no further study on the catalytic activity of platinum particles with smaller sizes although dramatic size-dependent catalysis is expected in this size region, probably because of difficulty of preparing and identifying such small particles accurately. To overcome this difficulty and unveil such sizespecific reactivity, a totally different approach should be contrived and introduced. In such an approach, it is desirable that the reaction be conducted on a size-selected platinum cluster or its ion, which is free of any environmental effects due to liquid media, solid supports, etc. Studies of this kind are realized in a collision reaction of a size-selected platinum cluster ion with a reactant molecule in the gas phase. Several groups have so far experimentally investigated reactions of Pt+ with hydrocarbons.6-14 Reactions involving platinum cluster ions, Ptn+, with small molecules also have been explored.13,15-18 These reactions have been performed mainly under conditions that the cluster ions collide with reactant molecules many times before the product ions are sampled, and in some cases the rate constants of the overall reactions have been obtained from the intensity decrease of the parent ion. For instance, Bondybey and co-workers have measured the rate constants of dehydrogenation of methane by Ptn+ (n ) 1-9) in a Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer; the rate constant generally increases and approaches to the collision rate as the cluster size increases, but has a minimum at the size of 4.13 Michl and co-workers have investigated dehydrogenation and cracking of n-butane by Pt+ and Pt2+ under a single collision condition in a triple-quadrupole * Corresponding author. E-mail: [email protected]. † East Tokyo Laboratory, Genesis Research Institute, Inc. ‡ Cluster Research Laboratory, Toyota Technological Institute.

Figure 1. Schematic view of the apparatus: (a) xenon ion beam, (b) platinum targets, (c) cooling cell, (d) octopole ion guides, (e) first quadrupole mass filter, (f) collision cell, (g) second quadrupole mass filter, (h) ion conversion dynode, and (i) secondary electron multiplier.

mass spectrometer.15 On the other hand, Morokuma and coworkers have calculated the geometric structures of platinum clusters (dimer and trimer) adsorbed with methane by using a density functional method.19,20 In the present study, we measured the absolute cross sections for the reactions of size-selected Ptn+ (n ) 1-5) with CH4, C2H6, C2H4, and C2H2, under a single collision condition, to investigate the initial catalytic processes induced by the collision. We have found that one hydrogen molecule is liberated by the collision (the initial step of dehydrogenation on Ptn+), and the dehydrogenation cross sections are extremely large and sizedependent. 2. Experimental Section The apparatus employed in the present study is described briefly in this report (see refs 21-23 for detailed descriptions). A schematic diagram of the apparatus is shown in Figure 1. The apparatus consists of a cluster ion source, a cooling cell, quadrupole mass filters, a collision cell, and a detector. Platinum cluster ions were produced by xenon-ion sputtering on four platinum targets mounted circularly with respect to the xenon ion beam produced from a plasma ion source24 (CORDIS Ar25/ 35c, Rokion Ionenstahl-Technologie); the energy of the xenon ion is increased up to 18 keV. The cluster ions were cooled in the first octopole ion guide filled with helium gas having a pressure of 10-3 Torr and a temperature of 300 K, and were

10.1021/jp021275z CCC: $22.00 © 2002 American Chemical Society Published on Web 11/05/2002

11466 J. Phys. Chem. A, Vol. 106, No. 47, 2002

Hanmura et al.

Figure 2. Mass spectra of ions produced by the collision of Pt2+ with (a) C2H4 and with (b) C2D4 at the collision energy of 0.15 eV. The peak assignment is given.

then transported into the first quadrupole mass filter for sizeselection of the cluster ions. A size-selected platinum cluster ion from the mass filter was admitted into the octopole ion guide surrounded with a collision cell in which reactions between the platinum ion and a reactant molecule under controlled collision energies take place. Commercially available reactant gases, CH4, C2H6, C2H4, C2D4, and C2H2, were used without further purification. A spinning rotor gauge (MKS, SRG-2) measured the pressure in the collision cell. To fulfill the single-collision condition, the pressure in the collision cell was maintained at ∼5 × 10-5 Torr. The product ions from the collision cell were mass-analyzed in the second quadrupole mass filter and were detected by a detector consisting of an ion conversion dynode and a secondary electron multiplier (Murata Ceratron, EMS6081B). The signal from the detector was processed in an electronic circuitry based on a personal computer. The spread of the translational energy of the parent cluster ions was determined to be typically 5 eV in the laboratory frame by applying a retarding voltage to the octopole ion guide mounted in the collision cell. This energy spread gives rise to the uncertainty of about (0.1 eV in the center-of-mass frame in a collision involving a platinum tetramer ion. The total reaction cross section, σr, is expressed as

kBT Ir + Σ Ip σr ) ln Pl Ir

(1)

where Ir and ΣIp represent the intensity of a parent ion passing through the collision region and the sum of the intensities of the product ions, respectively, P and T are the pressure and the temperature of a sample gas, respectively, l () 120 mm) is the effective path length of the collision region, and kB is the Boltzmann constant. 3. Results Figure 2 shows typical mass spectra of Pt2+(C2H2) and Pt2+(C2D2) produced by the collision of Pt2+ with C2H4 (panel a) and C2D4 (panel b) together with intact Pt2+ passing through the collision cell. The peaks of the mass spectra are broad because (1) the mass resolution is traded off with the peak intensity and (2) each peak consists of possible combinations of stable isotopes of Pt. It is evident from the mass spectra that dehydrogenation is the only process that takes place on Pt2+. The mass spectra of the other product ions show that all the hydrocarbon molecules studied undergo dehydrogenation in the entire size and collision energy range studied. It has been

Figure 3. Dehydrogenation cross sections of (a) CH4, (b) C2H6, (c) C2H4, and (d) C2H2 on Ptn+ at the collision energy of 0.15 eV as a function of the cluster size. The symbols (solid circles, solid squares, solid triangles, and solid diamonds) represent the experimental cross sections, and the solid lines are the calculated cross sections (see text). The error bars give one standard deviation of the cross sections experimentally obtained.

reported also that the dehydrogenation is the dominant process in the collision of a platinum ion and platinum cluster ions with hydrocarbons.6,11-14 The dehydrogenation is expressed as

Ptn+ + CH4 f Ptn+(CH2) + H2

(2)

Ptn+ + C2H6 f Ptn+(C2H4) + H2

(3)

Ptn+ + C2H4 f Ptn+(C2H2) + H2

(4)

Ptn+ + C2H2 f PtnC2+ + H2

(5)

The mass peaks were assigned by taking advantage of the fact that the width and the shape of a peak assigned to a given product ion are practically the same as that assigned to the corresponding parent ion. Then, the mass difference of the parent (Ptn+) and the product-ion peak gives the mass of an extra species attached to Ptn+. For instance, the product ions due to the reactions of Ptn+ with C2H4 and C2D4 are assigned to Ptn+(C2H2) and Ptn+(C2D2), respectively, because the masses for the peaks of the product ions are larger by 26 and 28, respectively, than that of the parent cluster ion (see Figure 2). Figure 3 shows the cluster size dependence of the cross sections for the dehydrogenation of CH4, C2H6, C2H4, and C2H2 at the collision energy of 0.15 eV in the center-of-mass frame; the cross sections were obtained from the parent- and the product-ion intensities by using eq 1. In the present experiment, the cross sections for the saturated hydrocarbon molecules are not sensitive to the cluster-size change except for methane, whose cross section at the cluster size of 2 is exceptionally high. On the other hand, the cross sections for the unsaturated

Dehydrogenation on Pt Cluster Ions

J. Phys. Chem. A, Vol. 106, No. 47, 2002 11467

Figure 4. Dehydrogenation cross sections for + C2H4 as a function of the collision energy. The solid circles exhibit the experimental cross sections, and the broken line shows the calculated cross sections (see text). The error bars give one standard deviation of the cross sections experimentally obtained. Pt3+

Figure 5. Schematic potential energy curve along the reaction coordinate for the dehydrogenation of CxHy on Ptn+. The energy minimums noted as A, B, and C correspond to the physisorbed, the chemisorbed, and the dissociatively chemisorbed states, respectively. See text for the definition of the energy barriers, ∆E1 and ∆E2, and the rate constants.

hydrocarbon molecules increase rapidly with the cluster size. Figure 4 shows the collision-energy dependence of the dehydrogenation cross section for Pt3+ + C2H4. The cross section decreases with the collision energy similarly to the cross sections for the other reactions studied in the present experiment. The present cross sections can be compared with those obtained by other groups. Armentrout and co-workers have measured the reaction cross sections for Pt+ + CH4 and CD4 collisions;14 the cross sections given in ref 14 are twice as large as ours at the collision energies less than 1 eV, probably because Pt+ used in the present experiment has a larger energy width. Bondybey and co-workers have measured the rate constants of the Ptn+ + CH4 reaction in a FT-ICR mass spectrometer.13 It is not possible to directly compare our data with theirs because ours are cross sections and theirs are rate constants. The cross sections were converted to our reaction efficiencies defined as the ratio of the reaction cross sections to the Langevin cross section (see Section 4 for the definition of the Langevin cross section) for the comparison with their reaction efficiencies. It was found that the size dependence of our reaction efficiency agrees fairly well with their reaction efficiency except for the size of 3; it is reasonable that the absolute values of ours and theirs do not agree because of the difference of the collision energy used.

the reactant system changes into the product system by surmounting the energy barriers of ∆E1 and ∆E2, lying along the reaction coordinate between the physisorbed and the chemisorbed states, and between the chemisorbed and the dissociatively chemisorbed states, respectively. The dehydrogenation cross sections were estimated by a procedure in which the physisorption cross section is approximated by the Langevin cross section for Ptn+ capturing CxHy in the charge-induced dipole field, and the subsequent reactions are given in the RRK scheme.25 Suppose that a physisorbed hydrocarbon molecule is either desorbed or chemisorbed by surmounting the energy barrier, ∆E1, with the rate constants of k1 and k2, respectively. Further, the chemisorbed hydrocarbon molecule either returns to the physisorbed one with a rate constant of k3 or proceeds to the dissociatively chemisorbed species with a rate constant of k4 by surmounting the energy barrier, ∆E2. It is assumed that the energetics determines the rate of the dissociatively chemisorbed species into the final product. The cross section of the physisorption is given by the Langevin cross section, σL,

4. Discussion

where R is the polarizability of CxHy (2.59 Å3, 4.47 Å3, 4.25 Å3, and 3.33 Å3 for CH4, C2H6, C2H4, and C2H2, respectively)26 and Ecol is the collision energy in the center-of-mass frame.27 The largest dehydrogenation cross section studied, which is observed for Pt4+ + C2H2 collision, is predicted well by the Langevin cross section, which should be the upper limit of the physisorption cross section. The cluster size dependence of the dehydrogenation cross section is introduced solely from the rate constants, k1, k2, k3, and k4 but is not from the Langevin cross section which does not depend on the cluster size, n. The dehydrogenation cross section is calculated from the rate equations for the physisorbed intermediate, the chemisorbed intermediate, and the dehydrogenated product based on the reaction scheme shown in Figure 5. In this scheme, the number density of the chemisorbed intermediate is considered to be sufficiently low that the steady-state approximation is applied. The reaction time (80 ∼ 250 µs) in the present experiment is sufficiently long that the exponential term turns out to be zero. Furthermore, it is considered that either k1 . k2 or k3 , k4 is true because neither the physisorbed intermediate nor the chemisorbed intermediate was observed. By solving the rate equations under these conditions, one obtains the cross section

It is conceivable, by analogy to the reaction of Pt+ with a methane molecule,6,8-10,12-14 that dehydrogenation proceeds as

Ptn+ + CxHy f Ptn+‚ ‚ ‚(CxHy) (physisorption)

(6)

Ptn+‚ ‚ ‚(CxHy) f [Ptn+CxHy]† (chemisorption)

(7)

[Ptn+CxHy]† f [Ptn+(CxHy-2)(H2)]† (dissociative chemisorption) (8) [Ptn+(CxHy-2)(H2)]† f Ptn+(CxHy-2) + H2 (H2 release) (9) Figure 5 shows a schematic reaction potential, which is simple enough to represent this reaction scheme. Initially, a hydrocarbon molecule, CxHy, is captured weakly on a platinum cluster ion, Ptn+, by a charge induced-dipole interaction (physisorption) and is then transferred to a chemisorbed state (chemisorption). The CxHy-chemisorbed platinum cluster ion changes into the final product ion via a platinum cluster ion with dissociatively chemisorbed species attached. Along the reaction coordinate,

σL ) π

( ) 2R Ecol

1/2

(10)

11468 J. Phys. Chem. A, Vol. 106, No. 47, 2002

Hanmura et al.

of the dehydrogenation, σ, as

k2 k4 σ ≈ × σL k1 + k2 k3 + k4

(11)

RRK theory gives the relations

(

k 1 ) A1 k 2 ) A2

(

)

Evib + Ecol Evib + Ecol + Ephys

N-1

(12)

)

Evib + Ecol + Ephys - ∆E1 Evib + Ecol + Ephys

N-1

(13)

where Evib, Ephys, Ecol, and ∆E1 represent the vibrational energy of the parent cluster ion considered to be (3n-6)kBT, the physisorption energy of CxHy on Ptn+, the collision energy, and the energy barrier between the physisorbed and the chemisorbed states, respectively (see Figure 5), A1 and A2 are the frequency prefactors related to the Ptn+-CxHy dissociation from the physisorption and the transfer to the chemisorption from the physisorption, respectively, and N is the total number of the vibrational modes of Ptn+CxHy less the internal modes of CxHy. The rate constants, k3 and k4, are given by

( (

) )

k 3 ) A3

Evib + Ecol + Ephys - ∆E1 Evib + Ecol + Echem

N-1

k4 ) A4

Evib + Ecol + Echem - ∆E2 Evib + Ecol + Echem

N-1

(14)

(15)

where Echem and ∆E2 represent the chemisorption energy of CxHy on Ptn+ and the energy barrier between the chemisorbed and the dissociatively chemisorbed states, respectively, A3 and A4 are the frequency prefactors related to the Ptn+-CxHy dissociation from the chemisorption and transfer to the dissociative chemisorption from the chemisorption, respectively, and N is the total number of the vibrational modes of Ptn+CxHy. Throughout the calculation, the prefactors, A3 and A4, are assumed to be the frequencies of the Pt-C stretching vibration (470 cm-1) and the C-H stretching vibration (2940 cm-1) of an ethylene molecule adsorbed on Pt(111) surface,28 respectively. Using eqs 10-15, one obtains the dehydrogenation cross sections at different collision energies at a given cluster size and internal temperature. Figure 4 shows the experimental (solid circles) and the calculated (broken line) cross sections for dehydrogenation of C2H4 on Pt3+ as a function of the collision energy. In this calculation, the dependence of the reaction cross section of Pt3+ on the collision energy observed was fit to that of the calculated cross section with three adjustable parameters, Ephys - ∆E1, Echem - ∆E2, and A1/A2; the best-fit values of Ephys - ∆E1, Echem - ∆E2, and A1/A2 were obtained to be -0.011 eV, -0.015 eV, and 0.33, respectively. It is likely that Echem - ∆E2 is most sensitively size-dependent because ∆E2 is related to the decomposition of the hydrocarbon molecule chemisorbed on Ptn+, which is sensitively influenced by the electronic structure of Ptn+. Therefore, the fitting for the other sizes was performed with a single adjustable parameter, Echem - ∆E2, while Ephys - ∆E1 and A1/A2 were fixed at the best-fit values obtained for n ) 3. Figure 3 depicts the cross sections thus calculated at the collision energy of 0.15 eV. Figure 6 shows the parameter, Echem - ∆E2, obtained in the calculation as a function of the cluster size. There is a strong correlation between the dehydrogenation cross section and the Echem - ∆E2

Figure 6. The parameter, Echem - ∆E2, for the dehydrogenation of (a) CH4, (b) C2H6, (c) C2H4, and (d) C2H2 as a function of the cluster size (see text for the definition of Echem and ∆E2). The errors were estimated from the uncertainties in the fitting.

value; the dehydrogenation cross section is large when the parameter, Echem - ∆E2, is large. In other words, the dehydrogenation proceeds more readily when the barrier height, ∆E2, is lower. It is not obvious whether CxHy at the chemisorbed state is molecularly chemisorbed as CxHy(a), chemisorbed as CxHy-1(a) + H(a), or otherwise. The calculations of the potential energy surface of the Pt+ + CH4 reaction have shown that the intermediate is expressed as H-Pt+-CH3.9,10,12,14,29 Morokuma and co-workers have computationally shown that CH4 on Pt2,3 is chemisorbed as CH3 and H on different Pt atoms.19,20 On the other hand, studies on the adsorption of C2H4 on Pt(111) surface30,31 have revealed that C2H4 is at first molecularly adsorbed as π-bonded species below 50 K and then changes into another adsorbate having two Pt-C σ bonds. Finally above 240 K, the adsorbed ethylene decomposes to ethylidyne (tC-CH3) on the surface. These findings suggest that the hydrocarbon molecules studied are also chemisorbed on Ptn+ as CxHy-1(a) and H(a), taking into account that the internal temperature of the chemisorbed intermediates rises as a result of the redistribution of the chemisorption energy. Seemingly, the larger barrier height, ∆E2, between the chemisorbed (Ptn+(CxHy-1)(H)) and the dissociatively chemisorbed states (Ptn+(CxHy-2)(H2)) is attributed to difficulty in the abstraction of two hydrogen atoms from CxHy on the cluster. Acknowledgment. This work was supported by Special Cluster Research Project of Genesis Research Institute, Inc. References and Notes (1) Bond, G. C. In Chemistry of the Platinum Group Metals: Recent DeVelopments; Hartley, F. R., Ed; Elsevier: Amsterdam, 1991; Chapter 2. (2) Sinfelt, J. H. In Catalysis: Science and Technology; Anderson, J. R., Boudart, M., Eds; Springer-Verlag: Heidelberg, 1981; Vol. 1, Chapter 5. (3) Che, M.; Bennett, C. O. AdV. Catal. 1989, 36, 55.

Dehydrogenation on Pt Cluster Ions (4) Masson, A.; Bellamy, B.; Colomer, G.; M’bedi, M.; Rabette, P.; Che, M. Proc. Int. Congr. Catal., 8th 1984, 4, 333. (5) Masson, A.; Bellamy, B.; Hadj Romdhane, Y.; Che, M.; Roulet, H.; Dufour, G. Surf. Sci. 1986, 173, 479. (6) Irikura, K. K.; Beauchamp, J. L. J. Phys. Chem. 1991, 95, 8344. (7) El-Nakat, J. H.; Dance, I. G.; Fisher, K. J.; Willett, G. D. Polyhedron 1993, 12, 2477. (8) Wesendrup, R.; Schro¨der, D.; Schwarz, H. Angew. Chem., Int. Ed. Engl. 1994, 33, 1174. (9) Heinemann, C.; Wesendrup, R.; Schwarz, H. Chem. Phys. Lett. 1995, 239, 75. (10) Pavlov, M.; Blomberg, M. R. A.; Siegbahn, P. E. M.; Wesendrup, R.; Heinemann, C.; Schwartz, H. J. Phys. Chem. A 1997, 101, 1567. (11) Taylor, W. S.; Campbell, A. S.; Barnas, D. F.; Babcock, L. M.; Linder, C. B. J. Phys. Chem. A 1997, 101, 2654. (12) Achatz, U.; Beyer, M.; Joos, S.; Fox, B. S.; Niedner-Schatteburg, G.; Bondybey, V. E. J. Phys. Chem. A 1999, 103, 8200. (13) Achatz, U.; Berg, C.; Joos, S.; Fox, B. S.; Beyer, M. K.; NiednerSchatteburg, G.; Bondybey, V. E. Chem. Phys. Lett. 2000, 320, 53. (14) Zhang, X.-G.; Liyanage, R.; Armentrout, P. B. J. Am. Chem. Soc. 2001, 123, 5563. (15) Magnera, T. F.; David, D. E.; Michl, J. J. Am. Chem. Soc. 1987, 109, 936. (16) Cox, D. M.; Fayet, P.; Brickman, R.; Hahn, M. Y.; Kaldor, A. Catal. Lett. 1990, 4, 271. (17) Kaldor, A.; Cox, D. M. J. Chem. Soc., Faraday Trans. 1990, 86, 2459.

J. Phys. Chem. A, Vol. 106, No. 47, 2002 11469 (18) Jackson, G. S.; White F. M.; Hammill, C. L.; Clark R. J.; Marshall, A. G. J. Am. Chem. Soc. 1997, 119, 7567. (19) Cui, Q.; Musaev, D. G.; Morokuma, K. J. Chem. Phys. 1998, 108, 8418. (20) Cui, Q.; Musaev, D. G.; Morokuma, K. J. Phys. Chem. A 1998, 102, 6373. (21) Ichihashi, M.; Hirokawa, J.; Nonose, S.; Nagata, T.; Kondow, T. Chem. Phys. Lett. 1993, 204, 219. (22) Hirokawa, J.; Ichihashi, M.; Nonose, S.; Tahara, T.; Nagata, T.; Kondow, T. J Chem. Phys. 1994, 101, 6625. (23) Ichihashi, M.; Hanmura, T.; Yadav, R. T.; Kondow, T. J. Phys. Chem. A 2000, 104, 11885. (24) Keller, R.; No¨hmayer, F.; Spa¨dtke, P.; Scho¨nenberg, M. H. Vacuum, 1984, 34, 31. (25) Draves, J. A.; Luthey-Schulten, Z.; Liu, W.-L.; Lisy, J. M. J. Chem. Phys. 1990, 93, 4589. (26) Lide, D. R. CRC Handbook of Chemistry and Physics, 78th ed.; CRC Press: Boca Raton, Fl, 1997. (27) Levine, R. D.; Bernstein, R. B. Molecular Reaction Dynamics; Oxford University Press: Oxford, 1974. (28) Ibach, H.; Lehwald, S. J. Vac. Sci. Technol. 1978, 15, 407. (29) Hada, M.; Nakatsuji, H.; Nakai, H.; Gyobu, S.; Miki, S. J. Mol. Struct. 1993, 281, 207. (30) Cremer, P. S.; Somorjai, G. A. J. Chem. Soc., Faraday Trans. 1995, 91, 3671. (31) Zaera, F. Langmuir 1996, 12, 88.